importing OME.tiffs in Omero 5

[yuriy_alexandrov]

...
on the glitch reported in the previous post -

it looks like local issue with server, as nobody can import anything there at the moment, but using "gretzky" server instead is fine.

Still the same problems remain, as above -

"ImageDescription" isn't displayed in Insight, but it is there in the OME-tiff file.
UUID FileNames put in the metadata are the same for every XY plane (the name of OME-tiff file), which is not as expected.
UUID FileNames put in ImageDescription are as expected - different for every XY plane.

Best,
Y.

[wmoore]

Hi Yuriy,

Just to let you know that the description XML issue is fixed:
http://trac.openmicroscopy.org.uk/ome/ticket/12056

[yuriy_alexandrov]

thanks Will - we tried to place xml generated by LabView into the Description field, and it comes out with <,> brackets and carriage return replaced by &lt; &gt; and #xD; characters, respectively.

Is this an expected behaviour? is there some other changes?

we are interested in saving custom xml in "Description".

Best,
Y.

[bpindelski]

Hi Yuriy,

When testing the "XML in image description" issue, I tried pasting in XML into the image description in Insight. There was no symbol substitution happening. To further investigate, could you upload a sample image that comes from LabView with an XML description, to our QA server (http://qa.openmicroscopy.org.uk/qa/upload/). Thanks.

Regards,
Blazej

[yuriy_alexandrov]

Thanks Blazej,

I uploaded a file - in both Matlab and Bio-Formats Viewer, the Description doesn't look like xml.

Just to clarify - I presume one needs to update Bioformats - where is the fix? -
is it here http://downloads.openmicroscopy.org/bio-formats/5.0.1/ ?

Is it Latest build or Daily build?

Best,
Y.

[bpindelski]

Hi,

I've tried importing your OME.TIFF to OMERO and viewing it in Insight. All node elements look formatted OK besides the end of each line, where I can see "#xD;". I'd suggest checking at the source of the XML, as this element is the carriage-return symbol (CR) but treated as a string literal instead of a HEX value. It might be missing an "&" from the beginning (i.e. "&#xD;").

Answering your BF question: the daily build is the unstable build used for testing new features, whereas the stable one is expected to be stable (more on that here: http://www.openmicroscopy.org/site/supp ... 5-0-latest).

Regards,
Blazej

[yuriy_alexandrov]

Hi again,

Thanks for looking at Description Blazej, I think it will be OK for our purposes, replaced symbols isn't big deal.
With "bigTiff" option now fixed, we were able to create OPT images >4G size, - importable by Omero Insight.

There is another issue however.
When I try to save tif-planes' filenames in the created OME-tiff's metadata, it doesn't work.
In the resulting file, they are all set to the image's filename, despite I use (very likely) the proper set of functions, for example -

Code: Select all

metadata.setUUIDFileName(sprintf(char(file_names{i})),0,i-1);   
metadata.setUUIDValue(sprintf(dicomuid),0,i-1);                                   
metadata.setTiffDataPlaneCount(toPosI(z),0,i-1);
...
writer.setMetadataRetrieve(metadata);

Another unresolved issue is adding Modulo annotation to the created OME-tiff.
It is quite relevant, as we have plenty of FLIM data where this Modulo specification is crucial for loading images into analysis software (first-foremost FLIMfit).

The modified version of "bfsave" is here
https://github.com/openmicroscopy/Imperial-FLIMfit/blob/master/FLIMfitFrontEnd/ic_importer/bfsave_with_description_and_UUIDFileNames.m

Best wishes,
Y.

[rleigh]

Hi,

Regarding storing the file each plane originated from, I think the recommended way would be to attach a FileAnnotation (or StringAnnotation) to the plane. components/formats-api/src/loci/formats/meta/MetadataConverter.java has examples of creating the annotations and then setting the plane reference with setPlaneAnnotationRef.

For the Modulo stuff, I'm not certain what the problem is here. If you've set the Modulo elements of CoreMetadata when before writing the OME-TIFF, the Modulo annotations should be automatically generated.


Regards,
Roger

[yuriy_alexandrov]

thanks for the suggestion Roger.

there might be quite a lot of these FileAnnotations, not sure how is that effective in database.

On Modulo - I've got example from Melissa how to do what You mentioned, in another thread.
http://www.openmicroscopy.org/community/viewtopic.php?f=6&t=7513

Best,
Y.

[yuriy_alexandrov]

Hi again,

It looks like there is problem with Omero-client "Export as OME-TIFF..." option (I looked at it in Insight).

We create Gated FLIM OME-tiffs with Modulo Annotation, that one can upload to Omero.
Then, these images might be loaded to our analysis program FLIMfit form there.
The OME-tiffs themselves might be loaded form disk by Bioformats version of FLIMfit that Ian developed (hopefully will be merged soon).

At the next turn, user should be able to download these images as OME-tiffs form Omero, without loss of metadata etc.

At least in Insight, this option doesn't work, throwing the error message "unable to export as OME-tiff".
Images were uploaded to Omero by the same Insight.
That was checked on 2 Omero-5 servers.

But maybe there is somewhere a version of Insight where this issue was fixed - if it is please let us know.

Best,
Y.